Homelands of the world’s language families

Abstract

A systematic, computer-automated tool for narrowing down the homelands of linguistic families is presented and applied to 82 of the world’s larger families. The approach is inspired by the well-known idea that the geographical area of maximal diversity within a language family corresponds to the original homeland. This is implemented in an algorithm which takes a lexicostatistically derived distance measure and a geographical distance measure and computes a lexical diversity measure for each language in the family relative to all the other related languages. The location of the language with the highest diversity measure is heuristically identified with the homeland.